Long-term plasticity of corticostriatal synapses is modulated by pathway-specific co-release of opioids through κ-opioid receptors

Abstract

Key points: Both endogenous opioids and opiate drugs of abuse modulate learning of habitual and goal-directed actions, and can also modify long-term plasticity of corticostriatal synapses. Striatal projection neurons of the direct pathway co-release the opioid neuropeptide dynorphin which can inhibit dopamine release via κ-opioid receptors. Theta-burst stimulation of corticostriatal fibres produces long-term potentiation (LTP) in striatal projection neurons when measured using whole-cell patch recording. Optogenetic activation of direct pathway striatal projection neurons inhibits LTP while reducing dopamine release. Because the endogenous release of opioids is activity dependent, this modulation of synaptic plasticity represents a negative feedback mechanism that may limit runaway enhancement of striatal neuron activity in response to drugs of abuse.

Abstract: Synaptic plasticity in the striatum adjusts behaviour adaptively during skill learning, or maladaptively in the case of addiction. Just as dopamine plays a critical role in synaptic plasticity underlying normal skill learning and addiction, endogenous and exogenous opiates also modulate learning and addiction-related striatal plasticity. Though the role of opioid receptors in long-term depression in striatum has been characterized, their effect on long-term potentiation (LTP) remains unknown. In particular, direct pathway (dopamine D1 receptor-containing; D1R-) spiny projection neurons (SPNs) co-release the opioid neuropeptide dynorphin, which acts at presynaptic κ-opioid receptors (KORs) on dopaminergic afferents and can negatively regulate dopamine release. Therefore, we evaluated the interaction of co-released dynorphin and KOR on striatal LTP. We optogenetically facilitate the release of endogenous dynorphin from D1R-SPNs in brain slice while using whole-cell patch recording to measure changes in the synaptic response of SPNs following theta-burst stimulation (TBS) of cortical afferents. Our results demonstrate that TBS evokes corticostriatal LTP, and that optogenetic activation of D1R-SPNs during induction impairs LTP. Additional experiments demonstrate that optogenetic activation of D1R-SPNs reduces stimulation-evoked dopamine release and that bath application of a KOR antagonist provides full rescue of both LTP induction and dopamine release during optogenetic activation of D1R-SPNs. These results suggest that an increase in the opioid neuropeptide dynorphin is responsible for reduced TBS LTP and illustrate a physiological phenomenon whereby heightened D1R-SPN activity can regulate corticostriatal plasticity. Our findings have important implications for learning in addictive states marked by elevated direct pathway activation.

Keywords: dopamine; dynorphin; electrophysiology; striatum; synaptic plasticity; voltammetry.

Figure 1. TBS LTP in striatal SPNs is modulated by light and transgene expression

A, TBS produces LTP in non‐light‐responsive (non‐LR) SPNs in the absence of 470 nm light, but TBS does not produce LTP during optogenetic activation of surrounding D1RCreChEYFP neurons. B, light‐responsive (LR) neurons do not exhibit TBS LTP in the absence of 470 nm light (‘dark’ condition), but optogenetic activation by 470 nm light transforms a slight LTD to a slight, but non‐significant LTP. For both A and B, inset shows representative EPSP traces prior to TBS (black) and 20 min after TBS (colour). Scale bar shows 5 mV and 10 ms. C, summary of the change in EPSP amplitude caused by TBS. ^** P < 0.005. Analysis of covariance with light × light‐responsiveness as categorical factors and current producing half‐maximal firing and mean spikes per burst as covariates was significant (F = 11.32, d.f. = 1, P < 0.0001). Post hoc contrasts show that (1) the PSP amplitude in response to TBS + light is significantly smaller than the response to TBS alone in non‐light‐responsive (non‐LR) neurons (F = 12.23, d.f. = 1, P = 0.0008), and (2) the PSP amplitude in response to TBS + light is significantly greater than the response to TBS alone in light‐responsive (LR) neurons (F = 10.2, d.f. = 1, P = 0.002).

Figure 2. Effect of cell excitability on plasticity outcome under dark (absence of 470 nm light) and 470 nm light conditions

A, relationship between plasticity outcome and current producing half‐maximal firing. Aa, dark conditions; Ab, light conditions. Correlation between plasticity outcome and current producing half‐maximal firing is significant only for the non‐light‐responsive (non‐LR) neurons in the presence of 470 nm light. B, relationship between plasticity outcome and number of spikes per burst during TBS. Ba, dark conditions; Bb, light conditions. Correlation between plasticity outcome and spikes per burst is significant only for the non‐light‐responsive (non‐LR) neurons in the dark. Note that during TBS in the presence of 470 nm light (Bb) the spikes per burst is significantly smaller for the light‐responsive neurons than for the non‐light‐responsive neurons.

Figure 3. Optogenetic impairment of LTP is rescued by the κ‐opioid receptor antagonist, but not the GABA_B receptor antagonist

A, NorBNI rescues LTP in non‐light‐responsive neurons in the presence of 470 nm light, without affecting the LTP response in the dark. B, NorBNI produces a slight increase in LTP in light‐responsive neurons in the presence of 470 nm light. C, CGP 35348 does not rescue the LTP deficit produced by 470 nm light. For A–C, inset shows representative EPSP traces prior to TBS (black) and 20 min after TBS (colour). Scale bar shows 5 mV and 10 ms. Lines without symbols in each panel show data from Fig. 1 for comparison. D, summary of the plasticity outcome for pharmacological groups in the light. ^* P < 0.05. Analysis of variance of three groups of light‐responsive neurons: absence of 470 nm light, presence of 470 nm light, and presence of 470 nm light + NorBNI, showed a significant difference (F = 3.36, d.f. = 3, P < 0.0229). Post hoc tests, corrected for multiple comparisons using Dunnett's, compared each group to the response in the presence of 470 nm light. The TBS response in the presence of 470 nm light + NorBNI was not significantly different (P = 0.912), but the response to TBS in the absence of 470 nm light was different from that of the 470 nm light group (P = 0.038). A separate analysis of variance of the 5 groups of non‐light‐responsive neurons showed a significant effect of light × drug (F = 2.58, d.f. = 4, P = 0.0432). Similar to the analysis of the light‐responsive neurons, post hoc tests were corrected for multiple comparisons using Dunnett's and compared each group to the response in the presence of 470 nm light. Both 470 nm light + NorBNI and absence of 470 nm light + NorBNI were significantly greater than the control (P = 0.049, P = 0.033, respectively). CGP 35348 does not change the response to TBS (P = 0.135). E, baseline EPSP amplitude is not affected by the KOR antagonist NorBNI (1 μm) in either LR (t = 0.87, d.f. = 22, P = 0.394) or non‐LR neurons (t = 0.07, d.f. = 26, P = 0.493).

Figure 4. Optogenetic activation of D1R‐SPNs reduces dopamine release via κ‐opioid receptor activation

A, inhibition of dopamine (DA) release by bath‐applied dynorphin. Traces show average and standard deviation for one experiment with 300 nm dynorphin, and subsequent rescue by NorBNI. B, summary bar graph showing the average reduction in dopamine release as a function of dynorphin concentration. For each dynorphin concentration, a paired t test compared the average baseline values and the average of the final two dynorphin A time points. C, average calibrated dopamine transients with representative subtracted current–voltage plot (inset) for experiments with and without concurrent optogenetic D1R‐SPN activation (n = 13 slices from 6 mice). D, average calibrated dopamine transients with representative subtracted current–voltage plot (inset) for experiments with and without concurrent optogenetic D1R‐SPN activation in the presence of the κ‐opioid receptor antagonist NorBNI (1 μm) (n = 7 slices from 5 mice). E, summary bar graph depicting the average decrease in dopamine release induced by optogenetic activation of D1R‐SPNs and the κ‐opioid receptor dependence of this effect. F, time course of effect of dynorphin A (DynA) on dopamine release, and rescue by NorBNI. ^** P = 0.005; ⁺⁺ P = 0.0073.

Figure 5. Cre expression in the Tg(Drd1a‐cre)EY217Gsat/Mmucd mouse line is enriched in striosomes

A, areas of dense μ‐opioid receptor (MOR) staining indicate the location of striosome compartments. B, EYFP illustrates non‐uniform Cre expression in dorsal striatum. C, overlay of A and B reveals colocalization of EYFP with striosomal compartments. Scale bars = 200 μm.